The 3 Best Artificial Intelligence (AI) Stocks for September

Nvidia’s Q2 fiscal 2027 revenue surged 106% YoY and management guided to 70% revenue growth for fiscal 2028, reinforcing bullish AI demand despite investor concerns. Micron benefits from a memory-chip supply shortage, with prices elevated and the company expecting new facilities to come online in 2027-2028 while still seeing tightness likely persisting before 2028; the stock trades around 6x fiscal 2027 earnings. Amazon’s AWS is accelerating (28% YoY growth in Q1 to 37% in Q2), with AWS generating 60% of operating income in Q2 (vs 21% of revenue), which the article argues should lift profits faster than revenue.

Analysis

AI spend is still cascading down the stack, but the most durable cash-flow surprise is likely in memory, not GPUs. As accelerators get denser, memory content per server rises faster than unit shipments, so MU can see margin expansion even if end-demand normalizes; that also squeezes server OEMs and storage vendors like DELL/HPE/WDC/STX via higher BOM costs and slower enterprise refresh cycles.

AMZN is the cleaner quality-growth expression because AWS operating leverage matters more than headline revenue. The near-term bear case is that capex depresses free cash flow before new capacity shows up in earnings, so the stock can stall even while fundamentals improve; the key watch item is whether AWS growth holds above the mid-30s through the next two quarters.

The contrarian point is that this is becoming a crowded AI-beta trade, so the best risk/reward may be relative value rather than outright longs. NVDA remains the toll collector, but the mispricing is whether the memory scarcity cycle lasts long enough for MU to re-rate before new supply arrives; if pricing rolls over early or hyperscaler demand pauses, the whole complex can de-rate quickly.

Key Decisions for Investors

• Long MU / short DELL or HPE for 3-6 months: best asymmetric pair if memory pricing stays tight; target 2:1 on continued DRAM/NAND price strength, stop if pricing or management commentary signals capacity relief before 2028.
• Buy AMZN on 3-5% pullbacks via a 3-6 month call spread: use it as a compounding AI infra position, with the trade invalidated if AWS growth slips back toward the low-20s or capex continues to outrun revenue conversion.
• Keep NVDA as a core AI hold, but prefer trimmed size or call-spread exposure rather than fresh full-size stock here: upside is still intact, but crowded ownership limits multiple expansion unless backlog and gross margin re-accelerate again.
• Watch for a rotation out of hardware beneficiaries into hyperscalers if server/OEM margins compress: if that starts showing up in next earnings season, reduce exposure to the broader semiconductor basket and lean into AMZN.